Accenture and Anthropic announced a partnership to establish embedded evaluators. They are reportedly expected to invest more than $1 billion in improving AI safety over the next 5 years.
Accenture and Anthropic announced on 5 October 2026 that they had partnered to establish a team of “embedded evaluators” responsible for AI model evaluation, red teaming, alignment evaluation and validation of safeguard effectiveness. The companies said they are expected to invest more than $1 billion in improving AI safety over the next 5 years.
The embedded evaluators will cover 4 areas: AI model evaluation, red teaming, alignment evaluation and validation of safeguard effectiveness. The evaluation team will work with Anthropic's divisions and operate from within the company, according to the announcement.
Accenture said it plans to advance the initiative by leveraging Faculty's expertise and combining practical knowledge in responsible AI with specialist capabilities in safety evaluation. Faculty is said to have experience evaluating AI across a broad range of fields, including government, defense and healthcare.
CTO Mark Warner said that Faculty's mission is based on the belief that AI safety should be achieved through design, and described its collaboration with Anthropic as embedded evaluators as an embodiment of the company's efforts.